There seems to be a malfunction in using Japanese words to look for the files whose descriptions are partly or entirely written in Japanese language.
If the searching keyword contains the same letters in the same order as those of the targeted Japanese phrase (without either spaces or any words in other languages between the Japanese letters), the results will be satisfactory because every Japanese context essentially never has any space letter in itself.
However, when the keyword just matches to some parts of the whole phrase, or when it contains not the suffix but only the body of a Japanese file name (e.g. not "あいうえお.jpg" but only "あいうえお"), it will never be successful for the wanted files to appear in the results.
I would like you to make it possible for the searching engine to recognize some fragmentary letters which are the parts of the entire Japanese phrase, and to identify Japanese file names without its suffixes, so that it will be a great help to all the users who have to discover files in Japanese.